🚀 LobeHub Release (20260528)

Release Date: May 28, 2026
Since v2.2.0: 220 merged PRs · 15 contributors

This cycle brings heterogeneous "platform agents" you can dispatch to local or remote devices, a rebuilt onboarding flow, document-centric chat, and a unified model-runtime error model — with new DeepSeek V4 and Gemini 3.5 Flash support along the way.

✨ Highlights

• More Hetero Agents (OpenClaw / Hermes) — Create heterogeneous agents and dispatch them to local or remote devices through the device gateway, with an execution-target switcher in the composer and persistent CLI sessions. (✨ feat(platform-agent): openclaw/hermes agent creation UI, device guard, and remote dispatch backend #15065, ✨ feat(hetero-agent): execution target switcher in composer #15179, ✨ feat(cli): integrate OpenClaw/Hermes hetero-agent dispatch with persistent sessions and notify protocol #15022)
• iMessage on Desktop — New iMessage setup and bridge on desktop, plus bot attachments across every platform. (✨ feat(bot): add iMessage Desktop setup and bridge #15228, ✨ feat(bot): add hidden iMessage backend foundation #15227, ✨ feat: support bot attachments across all platforms #15029)
• Skills in the Composer — Drag skill chips into chat, trigger installed skills from the slash menu mid-line, and surface project-level skills in the homogeneous agent runtime. (✨ feat(skills): drag skill chips + register agent-document skills #15095, ✨ feat(chat-input): add installed skills to slash menu with mid-line trigger #15061, ✨ feat(skills): recognize project-level skills in the homogeneous agent runtime #15110)
• New Models — DeepSeek V4 Flash/Pro and Gemini 3.5 Flash across providers, with thinking params for structured output and chat cost estimates. (✨ feat(opencode-go): align model configs with models.dev API, add DeepSeek V4 Flash/Pro, improve reasoning runtime #15031, 💄 style: add new Gemini 3.5 Flash model #15001, ✨ feat: support thinking params for structured output #15051, ✨ feat: add chat cost estimate support #14876)
• Agent Runtime Observability — OpenTelemetry GenAI semantic conventions plus per-call generation tracing. (✨ feat(observability): instrument Agent Runtime with OTel GenAI semantic conventions #15123, ✨ feat: per-call llm_generation_tracing observability #15124)

🤖 Agents & Heterogeneous Runtime

• Platform agent creation — OpenClaw/Hermes creation UI, device guard, and remote dispatch backend. (✨ feat(platform-agent): openclaw/hermes agent creation UI, device guard, and remote dispatch backend #15065)
• Execution-target switcher — Pick local vs remote execution directly in the composer; device-selection UX with actionable guidance. (✨ feat(hetero-agent): execution target switcher in composer #15179, ✨ feat(platform-agent): improve device selection UX with actionable guidance #15111)
• CLI hetero dispatch — OpenClaw/Hermes dispatch with persistent sessions and a notify protocol. (✨ feat(cli): integrate OpenClaw/Hermes hetero-agent dispatch with persistent sessions and notify protocol #15022)
• Gateway snapshot as source of truth — Consume the gateway uiMessages snapshot at step boundaries to keep chat state consistent. (✨ feat(chat): consume gateway uiMessages snapshot as SoT at step boundaries #15153, ✨ feat(agent-runtime): push UIChatMessage snapshot at gateway step boundaries #15152)
• Client sub-agent as a normal tool call — Simplifies the sub-agent execution path. (♻️ refactor(agent): run client sub-agent as a normal tool call #15281)
• Hermes agent chain — Implements the Hermes agent chain logic. (✨ feat(hermes): implement hermes agent chain logic #15189)
• Device registry — TRPC endpoints to register, list, update, and remove devices. (✨ feat(device): device registry TRPC (register / list / update / remove) #15299)
• Desktop device routing — Route gateway agent runs through lh hetero exec; restore userId in gateway dispatch and gate local-system by execution target. (🐛 fix(desktop): route gateway agent runs through lh hetero exec #15132, 🐛 fix: restore userId in gateway dispatch, gate local-system by executionTarget, add device switcher for regular agents #15232)
• Agent signals — Anchor agent-signal receipts to messages and isolate memory-agent messages into a child thread. (🐛 fix(agent-signal,app): anchor agent signal receipts to messages #14969, 🐛 fix(agent-signal): isolate memory-agent messages into a child thread #14921)

🚀 Onboarding

• Simplified first screen — Defer topic creation to first send. (✨ feat(onboarding): simplify first screen and defer topic creation to first send #15090)
• Market Agent Picker — Added as a classic onboarding step, with template prefetch. (✨ feat(onboarding): add Market Agent Picker as a classic onboarding step #14980, ✨ feat(onboarding): prefetch agent marketplace templates #15041)
• Welcome guidance — Show agent welcome guidance on first run. (✨ feat(onboarding): show agent welcome guidance #15098)
• Mobile — Adapt agent onboarding UI and restore Classic-step padding on mobile. (✨ feat(onboarding): adapt agent onboarding UI for mobile #15019, 🐛 fix(onboarding): restore mobile padding on Classic steps #15032)
• Discovery — Streamline discovery to a single profession question. (♻️ refactor(onboarding): streamline discovery to a single profession question #14987)
• Analytics — Track onboarding step events and create-agent modal source. (✨ feat(analytics): track onboarding step events #15133, ✨ feat(analytics): track create agent modal source #15028)

📄 Documents, Pages & Knowledge

• Thread chat in preview — Embed thread chat in the document preview portal. (✨ feat(portal): embed thread chat in document preview portal #15216)
• Non-markdown rendering — Render non-markdown docs as a read-only highlight. (✨ feat(agent-document): render non-markdown docs as readonly highlight #15272)
• Multi-select — Multi-select delete in the document tree. (✨ feat(agent-explorer): support multi-select delete in document tree #15125)
• Page-agent streaming — Preview initPage streaming arguments. (✨ feat(page-agent): preview initPage streaming arguments #15039)
• Per-agent topics — Per-agent topic management page. (✨ feat(topics): add per-agent topic management page #15207)
• Server-side category — Derive document category server-side and drop frontend predicates. (♻️ refactor(agent-document): derive category server-side, drop frontend predicates #15076)

🧩 Skills & Tools

• Drag skill chips — Drag skills into chat input and register agent-document skills. (✨ feat(skills): drag skill chips + register agent-document skills #15095)
• Slash menu — Installed skills appear in the slash menu with a mid-line trigger. (✨ feat(chat-input): add installed skills to slash menu with mid-line trigger #15061)
• Project skills — Recognize project-level skills in the homogeneous agent runtime and surface them regardless of active device. (✨ feat(skills): recognize project-level skills in the homogeneous agent runtime #15110, 🐛 fix(agent): surface projectSkills regardless of activeDeviceId #15177)
• VFS archiving — Archive oversized tool results to VFS instead of truncating. (✨ feat(tool): archive oversized tool results to VFS instead of truncating #15074)
• @localfile mentions — Drag folders into chat input as @localFile mentions on desktop. (✨ feat: drag folders into chat input as @localFile mentions on desktop #15071)

🧠 Model Runtime & Providers

• Error spec registry — Unify error codes into a spec + pattern registry, split ProviderBizError into finer codes, classify Cloud-only codes via a tier digit, and add DatabasePersistError. (♻️ refactor(model-runtime): unify error codes into spec + pattern registry #15262, ✨ feat(model-runtime): split ProviderBizError into finer codes + reclassify catch-all at write time #15286, ✨ feat(model-runtime): classify Cloud-only error codes via numericId tier digit #15278, ✨ feat(model-runtime): add DatabasePersistError code for failed DB queries #15279)
• New models — DeepSeek V4 Flash/Pro (opencode-go) and Gemini 3.5 Flash; DeepSeek V4 Pro on SiliconCloud. (✨ feat(opencode-go): align model configs with models.dev API, add DeepSeek V4 Flash/Pro, improve reasoning runtime #15031, 💄 style: add new Gemini 3.5 Flash model #15001, 💄 style: add Gemini 3.5 Flash to LobeHub provider #15017, 💄 style: add DeepSeek V4 Pro to SiliconCloud model list #15267)
• Structured output — Thinking params for structured output, Bedrock structured generation, and DeepSeek generateObject tool choice. (✨ feat: support thinking params for structured output #15051, 🐛 fix: support Bedrock structured generation #15174, 🐛 fix: support DeepSeek generateObject tool choice #15054)
• Cost — Chat cost estimate support; preserve usage cost in custom streams. (✨ feat: add chat cost estimate support #14876, 🐛 fix(model-runtime): preserve usage cost in custom streams #15218)

💬 Chat & User Experience

• Follow-up chips — Extend follow-up chip suggestions to general chat with scene-specific model config. (✨ feat(follow-up): extend follow-up chip suggestions to general chat #15101, ✨ feat(follow-up): allow scene-specific model config for follow-up action extraction #14797)
• Input drafts — Persist unsent input drafts across tab switches and prevent repeated draft restore. (🐛 fix(chat-input): persist unsent input drafts across tab switches #14992, 🐛 fix(chat-input): prevent repeated draft restore #15024)
• Command menu — Order topic/message search by recency and promote inline type filters. (🐛 fix(command-menu): order topic/message search results by recency #15094, 🐛 fix(command-menu): promote inline type filters from setSearch #14986)
• Zoom HUD — Show a zoom-level HUD on Cmd +/− and Cmd 0. (✨ feat(desktop): show zoom level HUD on Cmd+/- and Cmd+0 #15294)
• Copy — Unescape markdown escapes when copying user messages. (🐛 fix(copy): unescape markdown escapes when copying user messages #15253)

🖥️ Desktop

• App Nap fix — Prevent App Nap from dropping the gateway WebSocket during display sleep. (🐛 fix(desktop): prevent App Nap from dropping gateway WebSocket during display sleep #14994)
• File preview — Preview .cjs/.mjs/no-extension files instead of binary fallback and expand ~ when opening local files. (🐛 fix(desktop): preview .cjs/.mjs/no-ext files instead of binary fallback #15168, 🐛 fix(desktop): expand ~ when opening local files and folders #15284)
• Cross-platform settings — Open settings via main-window navigation on Windows/Linux and restore the route after an update restart. (🐛 fix(desktop): open settings via main window navigation on Windows/Linux #15036, 🐛 fix(desktop): restore route after update restart #14922)
• Token refresh — Prevent frequent logout from token-refresh retries. (🐛 fix(desktop): prevent frequent logout from token refresh retry #14928)

📊 Observability

• OTel GenAI — Instrument Agent Runtime with OpenTelemetry GenAI semantic conventions. (✨ feat(observability): instrument Agent Runtime with OTel GenAI semantic conventions #15123)
• Generation tracing — Per-call llm_generation_tracing with a pre-allocated tracingId and recordFeedback router. (✨ feat: per-call llm_generation_tracing observability #15124, ✨ feat(llm-generation-tracing): pre-allocate tracingId + recordFeedback router #15146)
• Error classification — Persist ERROR_CODE_SPECS classification on operation errors. (✨ feat(agent-runtime): persist ERROR_CODE_SPECS classification on operation errors #15273)

🗃️ Database Migrations

• Batch migrations — Topic usage stats, push tokens, tasks.editor_data, and document shares. (🔨 feat(db): batch topic usage stats, push tokens, tasks editor_data & document shares migrations #15280)
• Tracing & eval tables — Add llm_generation_tracing and agent eval experiment tables. (🔨 feat(db): add llm_generation_tracing and agent eval experiment tables #15126)

Self-hosted operators should run the database migration (pnpm db:migrate, or restart with auto-migrate enabled) after upgrading. The changes are additive and backwards-compatible.

🔒 Security & Reliability

• Security: Remove the getPlaintextCred tool to prevent plaintext credential exposure. (♻️ refactor(creds): remove getPlaintextCred tool to prevent plaintext credential exposure #14998)
• Security: Prompt account selection for Google OAuth and add prompt=consent to the OIDC authorization URL to fix missing refresh tokens. (🐛 fix(auth): prompt account selection for Google OAuth #15234, 🐛 fix(market-auth): add prompt=consent to OIDC authorization URL to fix missing refresh token #15010)
• Reliability: Preserve streamed content across a mid-stream cancel. (🐛 fix(agent-runtime): preserve streamed content across mid-stream cancel #15173)
• Reliability: Bound the Redis command timeout and configure the Anthropic client timeout. (🐛 fix: bound redis command timeout #15091, 🐛 fix: configure anthropic client timeout #15042)
• Reliability: Prevent infinite recursion in the assistant chain. (🐛 fix(conversation-flow): prevent infinite recursion in assistant chain #15288)

👥 Contributors

Huge thanks to 15 contributors who shipped 220 merged PRs this cycle.

@AnotiaWang · @sxjeru · @algojogacor · @hardy-one · @arvinxx · @Innei · @tjx666 · @lijian · @AmAzing129 · @rdmclin2 · @neko · @cy948 · @CanisMinor · @sudongyuer · @rivertwilight

Plus @lobehubbot and renovate[bot] for maintenance.

Full Changelog: v2.2.0...release/weekly-20260528